However, as things have changed I have now decided to move all my web activities over to the humongous and well known DreamHost hosting service. Moving it all has been extraordinarily boring, but at least it’s finished now so I can concentrate on making new, fun stuff for the web instead of worrying about my hosting, since I’ve known for about a month that I had to make this move. This three day weekend provided a good opportunity to move things over in stages to prevent a total meltdown of all my sites at the same time (which happened quite a lot when I hosted with Binero). That last remark brings me to why I’m changing to DreamHost. These are the reasons:

1. A more stable hosting service. Binero used to be very good and all, but at some point they just started sucking. This peaked last month when I found all my sites being broken due to changes made by Binero (without informing me in advance) three times in one month. That is just a bit too often for me to tolerate. I’ve had my sites broken by Binero in the past too, but not that often. And once it took them two weeks to fix a simple problem introduced by one of their changes, by which time I had of course already worked around the it.

Binero’s tech support has always been very quick to respond, but usually had some snide, unnecessary remark in the response as well, and usually blamed me for problems caused by them randomly changing stuff. I’ve been a customer with them for a long time, so even if they’ve been telling new customers how to set things up correctly for the last five years, that doesn’t mean I know it automatically without them telling me about it. It’s not good when being a loyal customer is used against you. So I definitely can’t recommend Binero to anyone who’s thinking about getting web hosting somewhere.

Lately they’ve started hyping a new system they call Binero 2.0, which is exactly what I don’t want – even more changes, bound to break stuff if I know Binero correctly. On the other hand, the sysadmin at work has had sites running at DreamHost for many years without any breakage or changes. That kind of stability is what I want. I also like DreamHost’s attitude that you can essentially do whatever you want as long as it’s not a security risk or malicious. Binero on the other hand disabled sftp (encrypted) access, allowing only ftp (unencrypted) access for security reasons, according to their tech support. That is about as stupid as it gets. Besides, they also disabled ssh shell access (of course, unannounced), which is a pain in the ass when you need to fix something quickly, besides being just silly.

2. Better located hosting for faster response times. When I started making web sites I lived in Sweden and most people who accessed the sites were in that area as well, so it made sense to use Sweden-based hosting. Now, however, I live in Japan and most of the people accessing my sites (whatlanguageisthis.com is by far the most accessed one now) are in North America, Japan, or India, as well as from all over Europe with no country in particular standing out. My ping time to my sites on Binero’s servers was 300 ms, but DreamHost has 125 ms ping, which is a very noticeable improvement. DreamHost’s servers are located in California, which is the center of the physical Internet, unlike Sweden, which is about as far out on the fringe as you can get except Russia and Africa. (For example, Internet traffic from Japan to Sweden goes via the US, not via Russia, and all Asian countries are connected via the US west coast.)

3. Get away from Sweden due to FRA. If you don’t know what FRA (“National Defence Radio Establishment”) is, would you believe me if I said that the Swedish government/military intercepts all Internet traffic crossing the border? Hopefully you wouldn’t, since you’re thinking that Sweden is a nice, modern, democratic society that doesn’t need to spy on its citizens nor the Russians, since the Cold War is over by now. Unfortunately, that’s what they’re doing, and Sweden seems to be moving in the wrong direction in many other ways too now when it comes to information privacy. Not nice. Since I’m in Japan, and stupid Binero didn’t allow encrypted access, that meant the FRA could intercept anything I uploaded to my sites. Not that I would upload something sensitive without encrypting it first, but still, it’s just silly. The US is a better place to host web sites, so score one more for DreamHost.

Anyway, that’s the end of this rant. Finally the move is over, and I’m looking forward to not having to wake up and find my sites being broken much often any more. I’ll also be writing more interesting stuff on this blog now that I’ve finished both the moves to a better blogging platform (WordPress instead of Blogger) and hosting service (DreamHost instead of Binero). No need to procrastinate any further now that the boring stuff is done.

I then took the test back in December. It felt like a pass, but I was unsure about my goal of getting a score over 80%. Anyway, today the result came:

So yeah, it’s a pass! And with a score of 84%, I consider it a success. It’s surely a relief to know that I won’t have to do another JLPT ever again.

Speaking of which, if I did a JLPT1 now, I’m sure I’d get a 90% score. Because I really got up to speed with my study routines before the test, and guess what – I never stopped! Yes, start at 100% and then increase, is the philosophy I follow. Since, in my opinion, I’ve now proved that I know how to study efficiently, I’ll keep writing about study methods and efficiently learning Japanese more from now on. I’ve got some good things going now, so I’m hoping that sharing my techniques can be of use to someone.

The score, as you can see, breaks down to 86% on writing/vocabulary, 80% on listening, and 85.5% on reading/grammar. So compared to the mock test I did before the real one, that’s much better on writing/vocab (up from 70%), slightly worse on listening (down from 82% – must have been that damned clown demon!), and a little better on reading/grammar (up from 81.5%). I’d speculate that the writing/vocab score was unreasonably low on my mock test for some unknown reason, and that reading/grammar went better on the real thing because I concentrate better when I know it’s for real, and 90 minutes is much longer than my usual attention span.

Low score on listening, as before, despite me listening to the radio while working, watching tv, and talking to my girlfriend, I cannot really explain. Actually, it might not have to do with Japanese in particular. I’m a good listener in the sense that I let people talk and make an effort to understand what they’re getting at, but I’ve realized I’m very bad at concentrating to someone talking for more than a minute or so… I lose concentration and start thinking of other things quite quickly. I’m the kind of person you want to send an email to rather than call on the phone if the matter requires more than 15 seconds to explain, if you see what I mean. Maybe that’s just it. Or maybe it’s because I had a very annoying, slight kind of ear disorder of my left ear the day of the test caused by a minor cold just before and sleeping without covering myself with the blankets properly. Even compared to my native northern Sweden, winter in Japan is a cold experience (since the houses are built by scammers and crooks).

Anyway, that’s it for today. Back to studying!

So how did the studying go? Well, decently good but not as good as I had hoped. For a while during fall I slowed down a bit (but never stopped) due to external issues. I kept reading books and listening to the radio, of course, to get continuous Japanese language input.

The weekend two weeks before the test, which was a three day weekend in Japan, I spent almost all days studying grammar. I was going through the Kanzen Master grammar book, reading every grammar item, the example sentences, and did all the exercises, in order basically.

Then I took the week before the test off, using my precious remaining paid holidays, in order to cram the last bits and pieces. Actually that was mostly vocabulary, but I also reviewed grammar, reading (the Unicom reading comprehension book), and listening (Unicom listening comprehension book).

I also entered the word lists I had gathered over the last year into Anki, and slashed the default intervals by at least a factor of 10 in order to cram the 755 words I knew I should but didn’t know. I had tried Anki before but was too appalled by the UI (I still am, and the fact that it’s slower than Java at starting up even on a dual core 2 gig machine), and a bit sceptical to using a computer for learning (yeah, feel free to not consider me Generation Y), but in the end I overcame this and it turned out pretty well, although I’d really recommend using Anki the way it’s intended to be used – which is as spaced repetition for long periods of time, not cramming. (It’s actually got a “cram mode” but I found that pretty useless – slashing the intervals proved to be better.)

Anyway, about a month before the test I did the writing/vocabulary part of a mock test and got pretty much the same result as before – even slightly lower – with 80% on kanji and 60% on vocabulary compared to 82% and 64% before. I don’t know why it was lower, maybe just random disturbance. However, when I did the rest of the mock test about a week before the real test, I had 82% on the listening (up from 72%), 74% on the reading (up from 68%), and a whopping 89% on the grammar (up from 78%). So the intense grammar studies had clearly paid off. Remember that the grammar section takes 20 minutes – 11% – of the test time, but still account for 25% of the score, and is the easiest section to cram.

So in total I had 78% on the second mock test – even though I did the vocabulary part before cramming vocabulary – so in the end, the forecast is looking good. When I took level 2 I had 65% on the mock test a week before, and ended up getting 81% on the real one, so I think I pull myself together when it’s for real too…

But I also think I was a little bit unlucky with some of the content that the real test covered. There were some topics appearing especially on the first part where my vocabulary is lacking. But even allowing for a 5% lower score due to that, it’s likely to be a pass, but may be closer to 70% than 80%, which I consider to be the lowest acceptable score. If I don’t reach that, I’ll probably do the test again soon.

Anyway, from this peroid of intensive study, I can at least draw these conclusions:

• Exercise books are good for self study. For some reason I had a lot of books, some of which I even read frequently, that included exercises, but I never did the exercises. The Kanzen Master books for instance are good, and the Unicom listening comprehension book. I guess if you follow their recipe of doing one chapter per day then you’ll be in good shape for the JLPT.

  Maybe it was because I was introduced to them as part of classes that I got off on a bad start. Doing exercises is good for your memory. I’ll definitely finish the kanji part of the Kanzen Master book too – I didn’t do that because my kanji skills are already good enough for the test.

• Use a computer program for vocabulary training! I have to admit I was being foolish not to do this from the start. After having used Anki for a few weeks now I realize how much simpler studying vocabulary has become. Especially using the synchronization feature I can keep my vocabulary synced between home and work (I often add work-related words). Also the fact that the software keeps track of which items need attention is very convenient.

So now we just have to wait for the score? No, now we keep on studying. All this studying has reminded me of how fun it is both to study and to learn, and not to mention the greatness of being able to communicate and read books in Japanese. Fortunately, there is still more to learn.

I’ve now done a mock test, using the Unicom book that contains two mock tests, to get a grasp of where I’m at and what I need to focus on. I can highly recommend that book, by the way. I used it for level 2 as well. Besides the tests, it assists in analyzing your weaknesses and tips on what you need to study.

Anyway, here’s a breakdown of my scores:

• Kanji: 82%
• Vocabulary: 64%
• Listening: 72%
• Reading: 68%
• Grammar: 78%

Interestingly, that means the average score for each of the three sections (kanji/vocabulary, listening, and reading/grammar) is 72% – quite a coincidence.

So what to make of this? First: it’s a pass, with a 288 p/72% score. That also means I’m on track for my goal to pass with more than 80%. When I do the mock test at home I’m more strict than at the real thing in that I don’t choose randomly when I don’t have a clue, and I try to finish it as fast as possible – I don’t stop to think and I don’t use spare time for reviewing.

I do that because I want data on how much time I actually need so that I can plan how much time to spend on the different parts during the real test – potential points vs time. For the reading section I had more than 15 min to spare, so I think this affects the end result by a few percentage points. Also, when I did the same thing for 2-kyu two years ago at this time, my score was barely above 60%, but on the real thing I scored 81%, so I think my score on the mock test is lower because I don’t concentrate as much as on the real test as well.

Second, the surprises: grammar score is high, reading is a bit low, and listening is lower than expected. I haven’t studied grammar really, but my studies consist mostly of reading, so I would have expected reading to be higher and grammar much lower. I felt very uncertain when answering many of the grammar problems even though I passed them. The only reason I can think of is that my book reading and radio listening have made me grasp grammar intuitively, much like a native speaker would.

The low reading score might be caused by me doing that section after coming home from work. I felt very tired by the end… And as I mentioned above I didn’t do any reviewing using spare time. After all, I read normal (actually, some of them are probably more academic than most people prefer to read) books written for native readers pretty much every day, and I don’t feel I’m missing out on the content of those books, so I don’t think my reading skill is bad. And time is definitely not a problem – my Japanese reading speed is good.

Low score on listening, despite listening to the radio for a few hours every day, I think was mostly caused by me not being up to date on the vocabulary used. Describing how people look and asking strangers for directions might be very common textbook examples, but it’s not something you do very often in real life… I am going to go through the Unicom listening comprehension book for 1-kyu as well, which contains the equivalent of about 4 tests’ worth of exercises, and that should be enough to easily get me above 80%.

Third, as expected: kanji is my strongest point and vocabulary is my weakest. Kanji are natural for me now, although recently I’ve been working on improving my kanji skills even more (I’ll write about my study methods some other day). But acquiring vocabulary is tough! I don’t really like repeating words or sentences or anything like that – I’m lazy – but I just hope to pick things up after seeing them enough times in books and news articles, and from hearing. The vocabulary used in JLPT is somewhat specific and specialized, albeit limited, and I have not been reading material specifically targeted at the test. Here as well, I am going to rely on the Unicom, namely reading comprehension book. But I’ll probably hold off on that until right before the test and keep reading normal literature that I enjoy reading for now.

Lastly for this post, I’d like to mention one more ingenious scheme I’ve come up with to extract more data from doing mock tests: marking certainty of the answers. I mark them essentially in 4 degrees, although I only make physical marks for 2: feel quite certain (no mark), feel a bit hesitant (one dot), feel like I’m mostly guessing (two dots), and don’t have a clue (no answer). Afterwards, I compile the percentage of correct answers for each certainty level (last level is obviously 0%). A stimulating paper exercise if there ever was one! But this time it also told me one thing: if I feel certain or hesitant doesn’t impact the score. But for the two-dot level the probability of a correct answer is halved. In other words I can go ahead and use my intuition even if I feel a bit hesitant, which saves time, and focus my reviewing (using time left after answering all questions) on a few questions that I felt very uncertain about.

Anyhow I’m interested in hearing about other’s progress on the JLPT and if you’re blogging about it, please post a link in a comment. Please also post comments on your own findings regarding the test. I’m quite exited about the test itself, besides becoming fluent in Japanese!

In the near future I also intent to write something about what I’ve learned about learning – because I feel I’m really getting into that now, and I’m already looking forward to the next language learning adventure – and also about my own study methods targeting JLPT1, and something about learning Japanese vs passing the JLPT.

Don’t forget to apply!